Skip to main content
Curb Cut Project

Curb Cut Project research, November 17, 2026

94.2% of local nonprofit homepages we tested have accessibility barriers that automated testing can detect

We tested the homepages of 70 nonprofit and community organizations in Southwest Riverside County against WCAG 2.2 Level AA. Organizations are not named.

of homepages had at least one detectable WCAG A or AA failure, at desktop size (64 of 70)
91.4%
had at least one failure rated critical or serious, at desktop size (64 of 70)
91.4%
different failed checks on the typical (median) homepage, at desktop size
3
affected page elements on the typical (median) homepage, at desktop size
10.5

The most common barriers

Share of the 70 tested homepages failing each check at least once, at desktop size, phone size or 320 pixels wide.

  1. Text with too little color contrast

    People with low vision or color blindness, and anyone reading a phone outdoors.

    75.7% (53 of 70)

  2. Links with no readable name

    Screen reader users hear only "link" and cannot tell where it goes.

    65.7% (46 of 70)

  3. Page scrolls sideways on small phones

    People who zoom in or use small phones have to scroll sideways to read every line.

    28.5% (20 of 70)

  4. Images without a text description

    Blind visitors using screen readers hear nothing, or a file name, instead of what the image shows.

    28.5% (20 of 70)

  5. Touch targets that are too small

    People with tremors or limited dexterity have trouble tapping small controls.

    25.7% (18 of 70)

  6. Embedded frames without a title

    Screen reader users cannot tell what an embedded map, video or form contains.

    14.2% (10 of 70)

  7. Pinch-to-zoom disabled on phones

    People with low vision cannot zoom in on phones.

    12.8% (9 of 70)

  8. Custom buttons and menu items with no readable name

    Screen reader users hear a control announced with no name, so they cannot tell what it does.

    11.4% (8 of 70)

  9. Links distinguishable only by color

    People who cannot distinguish colors cannot tell which words are links.

    7.1% (5 of 70)

  10. Hidden content that can still receive keyboard focus

    Keyboard users can land on content that screen readers say is not there.

    5.7% (4 of 70)

On a phone and at 320 pixels wide

On the same page load, each homepage was also checked at phone size (390 by 844 pixels, with touch emulation) and measured at 320 pixels wide, the width WCAG uses for zoomed-in reading.

had at least one detectable WCAG A or AA failure at desktop size, phone size or 320 pixels wide (66 of 70)
94.2%
had a failure that appeared only at phone size (12 of the 70 homepages checked at phone size)
17.1%
scrolled sideways at 320 pixels wide, so zoomed-in readers must scroll back and forth (22 of the 70 homepages measured)
31.4%

Failures found only at phone size

Share of the 70 homepages checked at phone size.

  • Touch targets that are too small5.7% (4 of 70)
  • Buttons with no readable name2.8% (2 of 70)
  • Links with no readable name2.8% (2 of 70)
  • ARIA attributes not allowed on the element1.4% (1 of 70)
  • Custom buttons and menu items with no readable name1.4% (1 of 70)
  • Hidden content that can still receive keyboard focus1.4% (1 of 70)
  • Invalid autocomplete values on form fields1.4% (1 of 70)
  • Images without a text description1.4% (1 of 70)
  • Form fields without a label1.4% (1 of 70)
  • Controls whose spoken name differs from their visible label1.4% (1 of 70)

How many problems a homepage has

Number of different automated checks each tested homepage failed, at desktop size. The typical (median) homepage failed 3.

Tested homepages by number of different failed checks
Failed checksHomepages
None detected6 (8.5%)
1 to 225 (35.7%)
3 to 533 (47.1%)
6 to 96 (8.5%)
10 or more0 (0%)

What fixing them looks like

Most of these are small changes in a website builder's settings. For the four most common:

Text with too little color contrast
Darken light gray or pastel text, or lighten its background, until the contrast ratio is at least 4.5:1 (3:1 for large text). Most website builders let you change this in the theme's color settings.
Links with no readable name
Give every link a readable name. Icon links (social media, search, menus) need hidden text or an aria-label such as "Our Facebook page".
Page scrolls sideways on small phones
Find the element wider than a phone screen (often a fixed-width image, table, embed or banner) and let it shrink to the screen width, for example max-width: 100%.
Images without a text description
Add a short description to each meaningful image in the image settings ("alt text"). Mark purely decorative images as decorative so screen readers skip them.

How we tested

The same 83 homepages of nonprofit and community organizations based in Temecula, Murrieta, Menifee, Lake Elsinore, Canyon Lake and Wildomar as the November 2026 study. One page load per homepage: axe-core 4.14 against WCAG 2.0/2.1/2.2 A and AA at desktop 1280x800, again at phone size 390x844 with touch emulation, and a sideways-scrolling (reflow) measurement at 320px wide. Nothing on the page was clicked, opened or typed into. robots.txt respected.

Each homepage was loaded once. On that one load, axe-core checked it against WCAG 2.2 A and AA rules in a desktop browser window (1280 by 800 pixels), then the same page was resized to phone size (390 by 844 pixels, with phone and touch emulation, without reloading) and checked again, and finally narrowed to 320 pixels wide to measure whether it scrolls sideways (WCAG 1.4.10, Reflow; up to 5 pixels is ignored). Nothing on the page was clicked, tapped, typed into or submitted, no keys were pressed and no menus were opened. The phone and 320-pixel checks shared a 20-second limit; when one could not finish, that homepage kept its desktop result and was left out of that figure, so each figure gives its own count.

Of 83 homepages on the list, 70 were tested. 13 could not be loaded. Percentages are of tested homepages. Testing engine: axe-core 4.14.1.

Earlier run and repeat check

This study replaces an earlier, desktop-only run on the same list of homepages: Southwest Riverside County nonprofit homepages, November 2026. That run found detectable failures on 64 of 70 tested homepages (91.4%). This study adds phone-size and narrow-screen checks, which is why it finds more.

We repeated the earlier run to check that the method gives stable results. Of the 65 homepages that loaded both times, 65 (100%) got the same verdict.

What these numbers do and do not mean

  • Automated testing finds only some accessibility problems. A homepage with no detected failures can still be hard to use, so the real share of sites with barriers is likely higher.
  • Only each homepage was tested, from one page load, at desktop size, phone size and 320 pixels wide. Other pages were not. Keyboard use and content inside closed menus were not tested, because we do not press or open anything on sites that have not asked us to.
  • These results are not a legal finding and do not mean any organization is out of compliance with a law.
  • Small organizations often run on volunteers and donated website builders. Many of these barriers take minutes to fix once someone knows they are there.

For reporters and researchers

Download the aggregate results (CSV). It contains the same numbers as this page and no organization names or addresses.

Download the per-homepage results (CSV). One anonymized row per homepage on the list: whether it was tested (or why not), how many checks it failed at each severity, and which axe-core rules failed. Anyone can recompute every figure on this page from it.

The data files, methodology and a script that recomputes every published figure are also in the public research repository on GitHub.

Suggested citation: Curb Cut Project, “Southwest Riverside County nonprofit homepages, desktop and phone, November 2026,” curbcutproject.org/research/swrc-nonprofit-homepages-phone-2026, published November 17, 2026.

Questions or interview requests: hello@curbcutproject.org

Run a nonprofit in the area?

The Curb Cut Project reviews nonprofit websites for free, explains each barrier in plain language and helps fix them. We do not share which organizations were in this study.

Questions about this research? About the Curb Cut Project